基于自适应高增益观测器的永磁同步电机预测电流控制方法

针对永磁同步电机(permanentmagnetsynchronous motor,PMSM)预测控制系统的模型失配、时变矩阵及观测器增益选择问题,提出基于自适应高增益观测器的预测电流控制方法(predictive current control based on an adaptive high-gain observer,AHGOPCC)。通过构建PMSM的前馈型扩张数学模型,集成考虑电机电阻、电感、磁链参数扰动项。提出基于前馈型扩张标称数学模型的高增益观测器,有效跟踪和估计电流与扰动状态及抑制转速变化引起的扰动。同时,设计自适应观测器增益矩阵满足全速工况的高增益条件,简化观测器的增益选择。在此基础上,构建离散化电流模型和成本函数,实现预测电流控制算法。实验结果证明了该方法的有效性,在多种工况下具备更优的快速性、鲁棒性及稳定性。     

Experimental investigation of the forced convective boiling heat transfer of R-600a/oil/nanoparticle

In this paper an experimental study of convective boiling heat transfer of R-600a/oil/nanoparticle mixtures is investigated. The experimental setup was prepared with a smooth horizontal tube as a test section with the length and diameter of 9.5 and 103 mm, respectively, and pure R-600a was applied for evaluating the heat transfer enhancement. Six mixtures containing 1% weight fraction of R-600a/oil with different concentrations of CuO nanoparticles including 0.0, 0.5, 1.0, 1.5, 2.0 and 5.0% weight fraction of R-600a/oil/nanoparticle were used in our study.The mass velocity per cross area was considered at the range of 50–700 kg/m² s for low vapor quality (ϕ < 0.25). The results showed that the convective boiling heat transfer coefficient will be increased by increasing the mass fraction of nanoparticles up to 2%, while by increasing the mass fraction of nanoparticles up to 5% the heat transfer coefficient will be reduced.     

基于机器学习的金属近表面缺陷超声检测方法

在超声脉冲回波检测技术中,由表面回波产生的"死区"会隐藏近表面缺陷的回波信号,对于近表面缺陷的实际检测造成一定的干扰.针对上述问题,提出一种将数字信号处理与支持向量机相结合的方法,通过识别缺陷一次回波和二次回波,实现对近表面缺陷位置的准确计算.实验中,通过垂直入射超声脉冲回波法采集轴承内圈的A波信号,使用支持向量机分类...     

Model based laser-ultrasound determination of hardness gradients of gas-carburized steel

Gas carburizing is a common industrial process utilized for case hardening of low carbon steels. However, there is a lack of non-destructive evaluation systems for the measurement of hardness-depth profiles. We propose a novel measurement method for the determination of hardness-depth profiles of two-step gas carburized steel specimens. The method is based on the measurement of broadband laser excited Rayleigh waves. Rayleigh waves were generated by a pulsed Nd: YAG laser in the thermoelastic regime and measured with a heterodyne Mach-Zehnder interferometer in the near-field. From two measurements with different source to receiver distances the dispersion diagrams were calculated by means of the phase spectral analysis method. In order to simulate the observed dispersive behavior of the Rayleigh waves, first the two-step gas carburizing process was simulated using solutions of the diffusion equation. The resulting continuous hardness profile was then discretized into up to 100 layers. Thereafter the Rayleigh wave dispersion diagram was calculated from the discretized stack of layers using a delta-matrix formulation of the Thomson-Haskell transfer matrix method. In order to obtain best fitting hardness profiles, the simulated dispersion diagrams were fitted to measurements with a curve fitting algorithm. Comparison of the Rayleigh wave inversion method with destructively obtained Vickers hardness profiles shows good quantitative agreement.     

Standby redundancy optimization problems with fuzzy lifetimes

Three types of system performances—the expected system lifetime, α-system lifetime, and system reliability—characterized in the context of credibility are investigated in this paper. Some fuzzy simulations are designed to estimate these system performances. In order to formulate general standby redundancy optimization problems with fuzzy lifetimes, a spectrum of standby redundancy fuzzy programming models are proposed. Fuzzy simulation, neural network, and genetic algorithm are also integrated to produce a hybrid intelligent algorithm for solving those models. Finally, some numerical experiments on multi-stage system and network system are provided.     

Comparison of Laguerre-Gaussian and Donut modes for MDM-WDM in OFDM-Ro-FSO transmission system

Radio-over-free-space-optics (Ro-FSO) technology may pave the way towards a ubiquitous platform for seamless integration of radio and optical networks without expensive optical fiber cabling. In this paper, to increase the capacity of Ro-FSO, mode division multiplexing (MDM) of two modes has been capitalized in a three-channel WDM system spaced by 1 nm over a FSO link of 80 km, resulting in a 120 Gbps six-channel Ro-FSO system. The SNR and received power of MDM of two Laguerre-Gaussian modes LG00 and LG01 is compared with respect to MDM of two transverse donut modes. The performance of four-level quadrature amplitude modulation (QAM) for orthogonal frequency division multiplexing (OFDM) of radio subcarriers in the WDM-MDMs system is investigated for mitigation of frequency-selective fading under strong atmospheric turbulence.     

A Straightforward Direct Traction Boundary Integral Method for Two-Dimensional Crack Problems Simulation of Linear Elastic Materials

This paper presents a direct traction boundary integral equation method (DTBIEM) for two-dimensional crack problems of materials. The traction boundary integral equation was collocated on both the external boundary and either side of the crack surfaces. The displacements and tractions were used as unknowns on the external boundary, while the relative crack opening displacement (RCOD) was chosen as unknowns on either side of crack surfaces to keep the single-domain merit. Only one side of the crack surfaces was concerned and needed to be discretized, thus the proposed method resulted in a smaller system of algebraic equations compared with the dual boundary element method (DBEM). A new set of crack-tip shape functions was constructed to represent the strain field singularity exactly, and the SIFs were evaluated by the extrapolation of the RCOD. Numerical examples for both straight and curved cracks are given to validate the accuracy and efficiency of the presented method.     

基于相位相关滤波与点云特征的激光点云配准方法

文中针对激光点云配准效率低和处理时间长的问题,提出一种基于相位相关滤波结合特征的扫描配准方法。扫描配准算法被解耦为旋转匹配与平移匹配两个步骤。在旋转匹配中,霍夫描述符提取的线段特征结合相位相关滤波得到具体的旋转量。平移匹配主要基于点云的边界特征,使用相位相关滤波处理提取到的特征,得到准确的x和y方向的平移量。与传统的迭代最近点(ICP)相比,提出的配准策略在处理静态数据时错误率降低了89.2%,处理时间降低了91.6%。同时,动态数据实验表明提出的方法具有较低的中位数和更好地一致性。